BƯỚC 1. DATA, VARIABLE CONSTRUCTION VÀ EMPIRICAL METHODOLOGY

Bài mẫu: Tourism Uncertainty and Corporate Dividend Policy
Working title: Tourism Uncertainty and Corporate Dividend Policy: International Firm-Level Evidence

Main X: Tourism Uncertainty Index (TUI), Ongan et al. (2026)
Main Y: Cash dividend policy
Financial data: Compustat Global + North America where needed

Critical design fact

Country-specific TUI is a country-time variable:

$$
TUI_{c,t}
$$

Therefore:

• Firm FE = possible;
• Year FE = possible;
• Industry×Year FE = possible;
• Country×Year FE = not possible for the main TUI coefficient;
• Country×Year FE becomes useful in an exposure interaction design:

$$
TUI_{c,t}times TourismExposure_i
$$

---

PHẦN 0. CORE PAPER PACK

Paper	Vai trò
Ongan, Gocer, Işık & Dogru (2026), Sustainable Development	Original TUI source and construction
Attig et al. (2021), Journal of Corporate Finance	International EPU–dividend evidence; higher EPU associated with higher payout
Adra et al. (2023), IRFA	GPR and payout choice; repurchases fall while cash dividends are insignificant
Choi & Park (2024), IRFA	EPU and dividend policy; private firms reduce payouts under uncertainty
Nguyen et al. (2026), Economic Modelling	National GPR increases cash dividends/payout in emerging economies
Li, Zhao & Zhang (2024)	China EPU associated with higher cash dividends
Kilincarslan & Demiralay (2021)	Dividend policy of travel and leisure firms; smoothing, profitability, debt, size
Kim & Jang (2009)	Financial characteristics of dividend-paying hospitality firms
Canina et al. (2001)	Dividend policy/signaling in lodging industry
Lintner (1956)	Dividend smoothing / partial adjustment
Fama & French (2001)	Who pays dividends
DeAngelo, DeAngelo & Stulz (2006)	Life-cycle / retained earnings
Brav et al. (2005)	Managerial payout decisions

0.1. Direct novelty warning

Do not claim:

“This is the first study of uncertainty and dividend policy.”

That literature is already large.

A defensible novelty claim is narrower:

• first/early use of Tourism Uncertainty Index in corporate payout research, subject to a final literature check;
• tourism-specific uncertainty rather than EPU/GPR;
• tourism-firm or tourism-exposure design;
• monthly fiscal-year-aligned TUI;
• cash dividend, payer decision, payout ratio, dividend cuts/omissions and total payout;
• cross-country tourism-dependence and firm-exposure heterogeneity.

Never use “first” without a final systematic search.

---

PHẦN A. MAIN X — TOURISM UNCERTAINTY INDEX

1. Source

Ongan, Gocer, Işık and Dogru (2026), Measuring Tourism Uncertainty (TUI).

The official TUI data are hosted by PolicyUncertainty.com and Mendeley Data.

Core properties:

• 43 countries;
• monthly series;
• starts in 2002;
• Context Window textual methodology;
• quarterly and annual aggregation;
• Global equal-weighted TUI;
• Global GDP-weighted TUI.

2. Preferred main TUI

Use:

$$
TUI_{c,t}
$$

country-specific TUI.

Why?

Corporate dividend policy is country/firm specific, so country-level TUI retains useful cross-country variation.

3. Global TUI

Use only as robustness.

If:

$$
GlobalTUI_t
$$

is common to all firms in year $t$:

full Year FE absorb its annual main effect.

Do not repeat the UCT error of estimating a pure time-series regressor together with full Year FE.

---

PHẦN B. ANNUAL VS FISCAL-YEAR-ALIGNED TUI

4. Calendar-year annual TUI

Simple baseline:

$$
TUI{c,t}
rightarrow DividendPolicy{i,c,t+1}
$$

5. Fiscal-year-aligned TUI

Preferred stronger measurement if monthly TUI and datadate are available.

For each firm-year:

$$
TUIFY_{i,c,t}

Average(TUI_{c,m})
$$

over the firm’s fiscal-year window.

Advantages:

• aligns uncertainty with the firm’s actual reporting/decision period;
• avoids assuming every firm’s fiscal year is Jan–Dec;
• creates timing that better matches dividend decisions.

Required diagnostic:

• fiscal-year-end month distribution;
• number of monthly TUI observations per firm-year;
• correlation annual TUI vs fiscal-year TUI.

---

PHẦN C. MAIN Y — DIVIDEND POLICY

6. Main cash dividend measure

Recommended:

$$
DividendAssets_{i,t}

frac{DVC{i,t}}{AT{i,t}}
$$

where DVC is the exact Compustat cash/common dividend field used in the data build.

Exact-code rule

Before paper writing, verify from Compustat dictionary:

• exact DVC definition;
• currency;
• consolidated treatment;
• zero vs missing.

Do not infer from variable abbreviation alone.

7. Dividend Payer

$$
DividendPayer_{it}

1(DVC_{it}>0)
$$

Zero dividend is a legitimate observation, not missing.

8. Dividend Payout Ratio

Possible:

$$
PayoutRatio_{it}

frac{DVC{it}}{IB{it}}
$$

But only use when:

$$
IB>0
$$

or under another clearly justified earnings denominator.

Do not interpret payout ratio mechanically for zero/negative earnings.

Report sample selection.

9. Dividend-to-Sales

$$
DVC/SALE
$$

Optional robustness.

10. Dividend Yield

If market value data are available:

$$
DividendYield

DVC/MVE
$$

exact scaling required.

11. Dividend Change / Cut

Important for uncertainty.

Possible:

$$
DividendGrowth

frac{DVCt-DVC{t-1}}{DVC_{t-1}}
$$

only where prior dividend is positive.

More robust event variables:

$$
DividendCut=1(DVCt<DVC{t-1})
$$

$$
DividendOmission=1(DVCt=0, DVC{t-1}>0)
$$

These are economically meaningful because dividends are sticky.

---

PHẦN D. TOTAL PAYOUT / REPURCHASES

12. Optional extension

If exact Compustat repurchase variable is validated:

$$
TotalPayout

Dividends + ShareRepurchases
$$

This is valuable because uncertainty may affect:

• flexible repurchases;
• sticky cash dividends

differently.

Adra et al. (2023) show GPR reduces repurchases but has insignificant cash-dividend effects.

Critical

Do not use PRSTKC as “share repurchase” without verifying its exact Compustat definition and construction.

---

PHẦN E. SAMPLE DESIGN

13. Design A — Tourism Firms Only

Preferred direct sample:

• Airlines;
• Hotels/Lodging;
• Restaurants;
• Travel/Tourism;
• possibly Gambling/Casinos/Recreation in broader definition.

Advantages:

TUI is directly economically relevant.

Disadvantage:

smaller sample and country coverage.

14. Design B — All Firms in TUI Countries

Estimate TUI main effect across all firms.

Interpretation is broader:

tourism uncertainty as a country-level uncertainty shock.

But economic relevance to non-tourism firms may be weak.

15. Design C — All Firms + Tourism Exposure

Strongest conceptual design:

$$
DividendPolicy_{i,c,t+1}

beta
(TUI_{c,t}times TourismExposure_i)
+
FirmFE
+
Countrytimes YearFE
+
Industrytimes YearFE
+
varepsilon
$$

where:

$$
TourismExposure_i=1
$$

for predetermined tourism-related firms/industries.

Country×Year FE absorb:

• TUI main effect;
• GDP;
• inflation;
• country uncertainty;
• policy shocks.

Identification asks:

Do tourism-exposed firms adjust dividends more strongly than non-tourism firms facing the same country-year environment?

This can be more credible than a tourism-only main-effect regression.

---

PHẦN F. TOURISM CLASSIFICATION

16. Core tourism sample

Recommended core economic groups:

• Airlines;
• Hotels/Lodging;
• Restaurants.

17. Broad tourism sample

Add:

• Travel services;
• Casinos/Gambling;
• Recreational services;
• Tourism operators.

Do not hardcode unexplained SIC/GICS

Create:

TOURISM_CLASSIFICATION_TABLE

Industry code	Industry name	Core/Broad	Tourism relevance	Keep?

The manuscript must map codes to economic categories.

---

PHẦN G. TIMING

18. Main model

Recommended:

$$
TUI{c,t}
rightarrow DividendPolicy{i,c,t+1}
$$

Why?

Dividend decisions often reflect:

• earnings;
• liquidity;
• uncertainty;
• expected financing needs.

A lead outcome improves temporal ordering.

19. Alternative contemporaneous model

$$
TUI{c,t}
rightarrow DividendPolicy{i,c,t}
$$

Use as robustness.

20. Persistence

Estimate:

$$
t+1, t+2
$$

if economically justified.

Dividend policy is sticky, so effects may adjust slowly.

---

PHẦN H. BASELINE MODEL

21. Main country-TUI specification

$$
DividendPolicy_{i,c,t+1}

beta TUI{c,t}
+
Gamma’FirmControls{i,t}
+
Delta’MacroControls_{c,t}
+
FirmFE_i
+
YearFEt
+
varepsilon{i,c,t+1}
$$

Expected sign?

Ambiguous.

This is fundamental.

---

PHẦN I. COMPETING THEORIES

22. Precautionary / liquidity-preservation channel

Tourism uncertainty can lower:

• demand visibility;
• expected cash flows;
• financing access.

Firms preserve cash:

$$
TUIuparrow
rightarrow Dividendsdownarrow
$$

23. Signaling / reassurance channel

Under uncertainty, managers may use stable/increased dividends to:

• signal confidence;
• reassure investors;
• reduce information uncertainty.

Then:

$$
TUIuparrow
rightarrow Dividendsuparrow
$$

24. Agency channel

Under uncertainty, shareholders may value cash distribution to reduce:

• free-cash-flow agency costs;
• managerial discretion.

Can imply higher payout.

25. Dividend smoothing channel

Firms strongly avoid dividend cuts.

Therefore:

TUI may first reduce:

• repurchases;
• dividend growth;

before reducing existing dividends.

This makes payer/cut/omission outcomes important.

---

PHẦN J. FIXED EFFECTS

26. Firm FE

Controls time-invariant:

• payout culture;
• ownership style;
• business model;
• country membership;
• long-run dividend propensity.

27. Year FE

Valid with country-specific TUI.

Controls global shocks.

28. Industry×Year FE

Useful in all-firm sample.

Controls sector payout cycles.

29. Country×Year FE

Absorbs:

$$
TUI_{c,t}
$$

main effect.

Do not use for direct TUI coefficient.

Use with firm-varying interaction:

$$
TUItimes TourismExposure
$$

---

PHẦN K. VCE / CLUSTERING

30. TUI varies at country-year

Therefore treatment dependence is country-level.

Preferred baseline:

cluster by country.

Report:

• number of country clusters;
• countries;
• years;
• firms.

If few country clusters:

• wild-cluster bootstrap;
• country leave-one-out;
• coefficient stability;
• alternative two-way inference.

Firm clustering alone is not sufficient for a country-level TUI shock.

---

PHẦN L. FIRM CONTROLS

31. Core controls

Recommended:

Size

$$
ln(AT)
$$

Profitability

Exact profitability metric.

Leverage

Debt/Assets.

Cash

$$
CHE/AT
$$

Growth Opportunities

Market-to-Book / Tobin Q.

Sales Growth

Retained Earnings

Dividend life-cycle variable:

$$
RE/TE
$$

or exact available construction.

Optional

• Tangibility;
• Cash Flow;
• Capex;
• firm age.

32. Dividend-policy theory requires controls

Particularly important:

• profitability;
• size;
• growth;
• leverage;
• cash;
• retained earnings.

Kilincarslan & Demiralay (2021) show travel/leisure dividend behavior varies with profitability, debt and size.

---

PHẦN M. MACRO CONTROLS

33. Baseline

• GDP Growth;
• Inflation.

Optional:

• policy rate;
• exchange-rate change;
• domestic credit;
• tourism growth;
• unemployment.

34. General uncertainty controls

Strong robustness:

• EPU;
• GPR;
• WUI;
• VIX/global financial uncertainty.

Purpose:

show TUI is not merely a generic uncertainty proxy.

But check correlations/multicollinearity.

---

PHẦN N. WINSORIZATION

Winsorize:

• continuous dividend outcomes;
• continuous firm controls;

at:

$$
1^{st}/99^{th}
$$

Do not winsorize:

• TUI;
• macro variables;
• payer/cut dummies.

---

PHẦN O. DIVIDEND-SMOOTHING MODEL

35. Lintner-style extension

Dividend policy is dynamic.

Generic:

$$
Dividend_{it}

alpha
+
rho Dividend{i,t-1}
+
beta TUI{ct}
+
…
$$

The lagged dividend captures smoothing/persistence.

Caution

Lagged Y + Firm FE can create dynamic-panel bias if T is short.

Use:

• long-panel FE sensitivity;
• bias-corrected approach;
• GMM only with full diagnostics.

Do not make GMM a ritual.

---

PHẦN P. CORE REFERENCES

Ongan, S., Gocer, I., Işık, C., & Dogru, T. (2026). Measuring Tourism Uncertainty (TUI). Sustainable Development. DOI: 10.1002/sd.71156.

Attig, N., El Ghoul, S., Guedhami, O., & Zheng, X. (2021). Dividends and economic policy uncertainty: International evidence. Journal of Corporate Finance, 66, 101785. DOI: 10.1016/j.jcorpfin.2020.101785.

Adra, S., Gao, Y., Huang, J., & Yuan, J. (2023). Geopolitical risk and corporate payout policy. International Review of Financial Analysis, 87, 102613. DOI: 10.1016/j.irfa.2023.102613.

Choi, Y. M., & Park, K. (2024). Economic policy uncertainty and dividend policy: Insight from private firms. International Review of Financial Analysis, 96, 103692.

Kilincarslan, E., & Demiralay, S. (2021). Dividend policies of travel and leisure firms in the UK. International Journal of Accounting and Information Management, 29(2), 324–344. DOI: 10.1108/IJAIM-09-2020-0144.

Canina, L., Advani, R., Greenman, A., & Palimeri, I. (2001). Dividend Policy in the Lodging Industry. Journal of Hospitality & Tourism Research, 25(1), 69–89. DOI: 10.1177/109634800102500106.

Lintner, J. (1956). Distribution of incomes of corporations among dividends, retained earnings, and taxes. American Economic Review, 46(2), 97–113.

Fama, E. F., & French, K. R. (2001). Disappearing dividends: Changing firm characteristics or lower propensity to pay? Journal of Financial Economics, 60(1), 3–43.

DeAngelo, H., DeAngelo, L., & Stulz, R. M. (2006). Dividend policy and the earned/contributed capital mix. Journal of Financial Economics, 81(2), 227–254.

Brav, A., Graham, J. R., Harvey, C. R., & Michaely, R. (2005). Payout policy in the 21st century. Journal of Financial Economics, 77(3), 483–527.

---

PHẦN Q. REQUIRED TABLES

1. TUI Data Note.
2. Country Coverage Table.
3. Tourism Classification Table.
4. Dividend Variable Definition Table.
5. Dividend Zero/Missing Audit.
6. Positive-Earnings Payout Ratio Sample Table.
7. Firm Controls Table.
8. FE/VCE Decision Matrix.
9. Literature–Method Matrix.
10. Fiscal-Year TUI Diagnostic.

---

PHẦN R. OUTPUT BƯỚC 1

Students submit:

1. exact TUI source/version;
2. exact Compustat dividend definitions;
3. tourism sample classification;
4. country coverage;
5. annual/fiscal-year TUI choice;
6. dividend outcome hierarchy;
7. FE/VCE decision;
8. literature evidence matrix;
9. Data & Methodology draft 1,800–2,500 words.

PHẦN S. CHECKLIST

☐ TUI source = Ongan et al. 2026.
☐ country-specific vs global TUI distinguished.
☐ 43-country/2002+ coverage documented.
☐ annual vs fiscal-year TUI documented.
☐ DVC exact Compustat definition checked.
☐ zero dividend not coded missing.
☐ payout ratio handles negative earnings.
☐ tourism classification table built.
☐ Firm FE + Year FE baseline.
☐ Country×Year not used for main TUI.
☐ country-level clustering addressed.
☐ core dividend controls included.
☐ TUI not winsorized.
☐ uncertainty sign treated as theoretically ambiguous.
☐ direct hospitality dividend literature included.
☐ no false “first uncertainty-dividend paper” claim.

Chuyển sang Bước 2: Baseline Results.